International Normalized Ratio Self-Management Lowers the Risk of Thromboembolic Events After Prosthetic Heart Valve Replacement

Transcription

1 International Normalized Ratio Self-Management Lowers the Risk of Thromboembolic Events After Prosthetic Heart Valve Replacement Thomas Eitz, MD, Soren Schenk, MD, Dirk Fritzsche, MD, Andreas Bairaktaris, MD, Otto Wagner, MS, Heinrich Koertke, MD, and Reiner Koerfer, MD Department of Thoracic and Cardiovascular Surgery, Heart and Diabetes Center North Rhine-Westphalia, Ruhr Universität Bochum, Bad Oeynhausen, Germany Background. Although prosthetic valves are durable and easy to implant, the need for lifetime warfarin-based anticoagulation restricts their exclusive usage. We investigated if anticoagulation self-management improves outcome in a single-center series. Methods. Between 1994 and 1998, 765 patients with prosthetic valve replacements were prospectively enrolled and randomized to receive conventional anticoagulation management by their primary physician (group 1, n 295) or to pursue anticoagulation self-management (group 2, n 470). A study head office was implemented to coordinate and monitor anticoagulation protocols, international normalized ratios (INR), and adverse events. Patients were instructed on how to obtain and test their own blood samples and to adjust warfarin dosages according to the measured INR (target range, 2.5 to 4). Results. Mean INR values were slightly yet significantly smaller in group 1 than in group 2 ( vs 3.0.6, p < 0.001). Moreover, INR values of patients with conventional INR management were frequently measured outside the INR target range, whereas those with anticoagulation self-management mostly remained within the range (35% vs 21%, p < 0.001). In addition, the scatter of INR values was smaller if self-managed. Freedom from thromboembolism at 3, 12, and 24 months, respectively, was 99%, 95%, and 91% in group 1 compared with 99%, 98%, and 96% in group 2 (p 0.008). Bleeding events were similar in both groups. Timerelated multivariate analysis identified INR self-management and higher INR as independent predictors for better outcome. Conclusions. Anticoagulation self-management can improve INR profiles up to 2 years after prosthetic valve replacement and reduce adverse events. Current indications of prosthetic rather than biologic valve implantations may be extended if the benefit of INR self-management is shown by future studies with longer follow-up. (Ann Thorac Surg 2008;85:949 55) 2008 by The Society of Thoracic Surgeons Current mechanical heart valve prostheses not only provide minimal pressure gradients with an adequate flow pattern but are also very durable and less apt to mechanical failure than previous generations of mechanical prostheses. The risk of thromboembolism, however, still accounts for most of the complications after mechanical heart valve replacements, with an incidence of 0.9% to 3.6% per patient year [1]. To avoid the need for lifetime anticoagulation and related adverse events, biologic valves are often chosen, although the risk of valve degeneration increases progressively over time [2]. Increasing pressure gradients across biologic valves contribute to pathologic ventricular hypertrophy [3] and predict valve degeneration [4]. Furthermore, about 25% to 35% of the biologic valves must be exchanged by 15 years of implantation [5, 6], and Accepted for publication Aug 21, Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29 31, Address correspondence to Dr Eitz, Klinik für Thorax- und Kardiovaskularchirurgie, Herz- und Diabeteszentrum NRW, Georgstr. 11, Bad Oeynhausen, 32545, Germany; teitz@hdz-nrw.de. this second major heart operation in older patients is associated with an increased mortality [7 9]. Because of their limited durability, biologic prostheses led to decreased survival of patients aged younger than 65 years [9]. Thus, cardiologists and cardiac surgeons in every-day practice are as yet left with the choice between the frangible biologic valves but no anticoagulation issues, or the perdurable mechanical prostheses with the permanent risk of anticoagulation-related adverse events. Indications for implantation of a mechanical valve can be extended if anticoagulation-related adverse events were reduced. In an effort to improve outcome after mechanical valve implantation, our institution began to investigate the strategy of anticoagulation patient selfmanagement in the early 1990s. An attractive indicator of oral anticoagulation is the international normalized ratio (INR), because this ratio is easily acquired and can be directly compared between clinical laboratories. In a number of reports, we previously showed that if our patients self-managed the INR, the incidence of adverse events was reduced [1, 10] and the survival after mechanical valve replacement was improved [11]. The current 2008 by The Society of Thoracic Surgeons /08/$34.00 Published by Elsevier Inc doi: /j.athoracsur

2 950 EITZ ET AL Ann Thorac Surg INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES 2008;85: Table 1. Patient Characteristics Variable a INR Self-Management (n 470) INR Conventional Management (n 295) Age, y Female 127 (27) 109 (37) Atrial fibrillation 131 (28) 94 (32) Valve position Aortic 386 (82) 255 (86) Mitral 106 (23) 63 (21) Prosthesis b CarboMedics 139 (30) 94 (32) St. Jude Medical 219 (47) 107 (36) Medtronic 112 (24) 94 (32) a Age is presented as means standard deviation; categoric data are given as number (%). b CarboMedics, Sorin Group, Austin, TX; St. Jude Medical, Inc., St. Paul, MN; Medtronic Inc, Minneapolis, MN. study compares the INR profiles between self-management and conventional management with directly relate INR management with outcome. Patients and Methods Patients Between 1994 and 1998, 765 patients were prospectively enrolled into this single-center study at Heart and Diabetes Center (Bad Oeynhausen, Germany). Inclusion criteria were the indication for implantation of mechanical heart valve prostheses in any position, the availability of INR values, and complete information on adverse events. The study was approved by the local Review Board, and informed consent was obtained from every participant. Patients were randomized at the time of operation to have the anticoagulation managed conventionally by their general practitioner (group 1) or to pursue INR self-management (group 2). Subsequent crossover to the other group was permitted, leaving 295 patients in group 1 and 470 patients in group 2. Key patient characteristics are summarized in Table 1. Patients in the group of INR self-management were younger, but more of them were men. There was no difference in the position of the valves, aortic or mitral, or both. The prostheses used were Medtronic Hall (Medtronic Inc, Minneapolis, MN), St Jude Medical (St Jude Medical Inc, St Paul, MN), and CarboMedics (CarboMedics Inc, Sorin Group, Austin, TX). at their own discretion. The CoaguChek device has been shown to provide adequate accuracy and reliability in prior studies [12, 13]. In addition, each device was tested after 6 months of usage. The rationale was to ensure that there was no need for any sort of calibration, yet INR levels were accurate compared with our certified laboratory. The results indicated that INR levels obtained by the CoaguChek device are nearly identical to those tested in our certified clinical laboratory (data not shown). The protocol of INR self-management has been previously reported in detail [14]. In short, patients receive instructions on how to use the CoaguChek device, evaluate the INR readings, and adjust warfarin dosages as early as on postoperative days 4 to 10. In particular, recommendations are given with respect to adjustments of warfarin dosages if the INR reading is outside the target range. The patients must understand that adjustments of warfarin dosages cannot be predefined because of varying needs for vitamin K antagonism. They are also taught that dosage adjustments must consider previous dosages and INR profiles. The instructors ensure that every patient gained proficiency on these important aspects of INR self-management upon discharge from our institution. In addition, booster instruction courses are held at 6 months after valve implantation, during which the INR profile is evaluated, and additional recommendations are given as needed. INR values and questionnaires about adverse events were submitted to the Study Head Office by phone and facsimile. The Head Office was available to our patients 24 hours a day, 7 days a week, to provide medical advice as needed. All patients were required to return to our institution for an outpatient cardiologic checkup at 6, 12, 18, and 24 months. The common closing date of the study was 2 years postoperatively. Data Analysis Data are presented as mean (SD) or 68% confidence limits, as appropriate. Grade III thromboembolic or International Normalized Ratio Management and Follow-Up The anticoagulation of patients in group 1was managed conventionally by the general practitioner. In general, blood samples were acquired at the discretion of the practitioner (commonly once or twice month) to measure the INR at the local laboratory. Dosage adjustment was made accordingly by the practitioner. Patients in group 2 measured the INR by the CoaguChek device (Roche Diagnostics GmbH, Mannheim Germany) Fig 1. International normalized ratio (INR) values stratified to ranges below the target range ( 2.5), within the target range (2.5 to 4.5), and above the target range ( 4.5) in patients with conventional (Conv l, patterned bar) and self-management (Self- INR, filled bars). (*p )

3 Ann Thorac Surg EITZ ET AL 2008;85: INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES bleeding events were evaluated according to Karnofsky [15]. Time-related events were analyzed by nonparametric Kaplan-Meier and parametric hazard function methods [16]. The hazard package was obtained from the Heart & Vascular Institute at The Cleveland Clinic Foundation [17] and run on SAS 9.1 software (SAS Institute Inc, Cary, NC). Risk factors for thromboembolism or bleeding after valve replacement were identified by hazard function regression using variables including gender, age, body surface area, atrial fibrillation, left atrial, and enddiastolic ventricular diameter, ejection fraction, brand and position of the valves, the INR management, and the mean INR level per patient. 951 Results A total of 46,855 INR values were collected in 765 patients during 1291 patient-years of follow-up. The number of Fig 3. Freedom from adverse events, including thromboembolism (TE) and bleeding, after heart valve replacement. Squares represent percentages of patients without adverse events at the time after valve replacement (noted on horizontal axis), and the vertical bars are asymmetric 68% confidence intervals for these estimates (equivalent to 1 standard error). Numbers in parenthesis indicate patients free of adverse events. The solid line, enclosed within dashed 68% confidence limits, is a parametric estimate; from this is derived the risk of adverse events (hazard function) by month shown in the inset. INR values differed considerably between group 1 (conventional INR management) and group 2 (INR selfmanagement). Whereas group 1 had a mere 8005 INR values ( measurements per patient), group 2 accumulated 38,850 INR values ( measurements per patient; group 1 vs 2, p 0.001). The INR levels and profiles were compared. Mean INR values in group 1 were slightly, yet significantly smaller than in group 2 ( vs , p 0.001). Although both groups were targeting an INR range between 2.5 to 4.5, group 2 had significantly more values inside the range. Whereas 35% of INR values in group 1 were outside the target range, only 21% of those in group 2 did not remain within (p 0.001). Specifically, group 2 had fewer INR values below the therapeutic range (18% vs 33%, p 0.001, Fig 1). The scatter of INR values also differed considerably. For instance, the mean variance of INR values between two consecutive INR measurements was smaller in group 2 than in group 1 (0.346 vs 0.388, p 0.05). Figure 2A depicts percentages of changes in absolute INR values stratified for various INR ranges. Small changes in INR values (less than 0.3 and 0.3 to 0.6 units) were observed Table 2. Risk Factors for Thromboembolism or Bleeding After Valve Replacement Parameter Coefficient SE p Value Fig 2. (A) Change of international normalized ratio (INR) values between two consecutive measurements in patients with conventional INR management (Conv l, patterned bar) and self-management (Self-INR, filled bar.) (** indicates p 0.001). (B) Change of INR values in relation to times of no adverse event vs times of 4 weeks before an adverse event (AE) of thromboembolism or bleeding. Conventional INR management Higher mean INR Lower BSA Double valve replacement BSA body surface area; INR international normalized ratio; SE standard error.

4 952 EITZ ET AL Ann Thorac Surg INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES 2008;85: Fig 4. Freedom from adverse event effect of self-managed international normalized ratio (INR) management (Self) vs conventional INR management (Conv l), as derived from the hazard function. Other related indicators were fixed during the modeling process (that is, parameters were specified/set/entered into regression equation of hazard model) (INR, 2.9; body surface area, 1.89; single-valve replacement). The dashed lines represent 68% confidence limits. more often in group 2 than in group 1 (p 0.001). This could be a correlate of our clinical observation that patients with high awareness of INR levels tweak their warfarin dosage to remain exactly within their INR target range. In contrast, large INR changes of more than 0.9 units were seen more often in the conventional INR management group (Fig 2A), which could indicate counteracting an INR level off the target range. Thromboembolism and bleeding during the 2 years of follow-up was investigated. Freedom from thromboembolism at 3, 6, 12, 18, and 24 months, respectively, was 99%, 97%, 95%, 93%, and 91% in group 1 compared with 99%, 98%, 98%, 97%, and 96% in group 2 (p 0.008). In the group with conventional INR management, 21 patients (7.12%) had a thromboembolic event compared with 14 (2.98%) in the group of INR self-management. In contrast, bleeding events were similar in both groups with respect to absolute occurrence of 6.8% for group 1 vs 6.8% for group 2 and rate of occurrence at 3, 6, 12, 18, and 24 months, respectively, of 98%, 97%, 94%, 93%, and 93% for group 1 vs 99%, 98%, 96%, 94%, and 93% for group 2 (p 0.653). Figure 3 shows the nonparametric and parametric estimates for freedom of both adverse events as well as the decomposed hazard phase. The risk of any adverse event was higher in the first months after valve replacement, but then declined (Fig 3 inset). Multivariate analysis revealed conventional anticoagulation management to be the strongest risk factor for adverse events (Table 2). Additional factors included higher mean INR per patient, lower body surface area (BSA), and double valve replacement. Interestingly, neither atrial fibrillation nor brand and position of the valves were identified as significant risk factors in our model. As an example for a given clinical scenario in a patient with single valve replacement, a BSA of 1.89, and a mean INR of 2.9, all risk factors were entered into the regression equation and solved for the individual risk. As shown in Figure 4, freedom of any adverse events is significantly lower if the patient chose conventional INR management rather than INR self-management (85% vs 92% at 2 years follow-up). Finally, the change of consecutive INR values was related to the time of adverse events. During times of no adverse events, there were frequent, yet slight adjustments made to the INR levels (INR change of less than 0.3 and 0.3 to 0.6, Fig 2B). As already outlined, patients with INR self-management more often fine-tuned their INR profiles, and these patients commonly remained without thromboembolic complications during followup. In contrast, during the 4 weeks before an event, there were fewer such fine adjustments (Fig 2B), suggesting that patients either did not try to adjust their INR levels or that the proportion of coarse INR adjustments (ie, 0.9 units) was increased relative to small adjustments. In our experience of several thousand patients with INR selfmanagement, we have found many instances in which our patients try to reverse an INR level that is far below the target range by ingesting large amounts of warfarin. These patients often reveal a very uneven INR profile and are likely to experience thromboembolic events. Comment As demonstrated quantitatively by this study, INR selfmanagement results in stable INR profiles within the therapeutic range and with lower scatter as well as a decreased incidence of adverse events up to 2 years after prosthetic valve implantation compared with conventional INR management. Smaller deviations of the INR profiles were predominantly observed in times of no adverse events. Common risk factors for thromboembolic and bleeding events in heart valve recipients include (1) valve-related factors such as make and position, (2) comorbidities such as atrial fibrillation, coronary disease, and left ventricular dysfunction, and (3) other patient-related factors such as age, sex, and history of smoking [18, 19]. The current study found lower BSA to be related with adverse outcome, which could be a surrogate of female sex; however, sex itself was not a risk factor. Most of the other factors were not found to be significantly associated with thromboembolism or bleeding in this study. For instance, Ruel and associates [19] found a tilting aortic valve to be associated with adverse outcome. Although the tilting valve in our study, the Medtronic Hall valve, tended to be associated with more adverse events, this trend did not reach statistical significance. We did not find mitral valve replacement to be a risk factor in the current study. Unlike Hamamoto and colleagues [20], we found double valve replacement to be a risk factor, which commonly included mechanical prostheses in mitral position. The quality of oral anticoagulation in the current study may partly explain the aforementioned differences of risk factors in other reports. It is well known that adjusteddose warfarin anticoagulation excels fixed-dose treat-

5 Ann Thorac Surg EITZ ET AL 2008;85: INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES 953 ment in prosthetic valve recipients [21] or high-risk patients with atrial fibrillation [22]. All patients of our clinical study had adjusted-dose warfarin anticoagulation, and they were also very closely followed up and treated. It is therefore not surprising that, for instance, atrial fibrillation was not a risk factor in our model. Sufficient anticoagulation in most of our patients could have abolished another risk factor in this study, that is, age. In support of this, we show in subsequent clinical studies of INR self-management [23] that older patients did not experience more thromboembolic events than those of the current study, which were younger. Furthermore, the quality of oral anticoagulation may even be the strongest predictor of thromboembolism or bleeding by itself [24]. Whereas both groups in the current study were in very close adherence to INR target ranges with relatively few outliers, the group of INR self-management had better INR profiles. Consequently, this group had less anticoagulation-related events, and non-inr self-management was thus a risk factor for adverse outcome. It is important to note that the improved quality of oral anticoagulation by INR self-management was likely due to the abundance of INR values in each patient. We would expect the primary physician to have adjusted warfarin dosages as adequately as our patients if the physicians had the same amount of data available. Because they measured the INR values only once or twice a month, however, the warfarin dose adjustment may frequently have been to counteract out-of-range INR values. In contrast, patients pursuing INR selfmanagement adjusted their warfarin dosages every 2 to 3 days, which was likely to result in stable INR profiles. Thus, this massive amount of INR data provided the basis for therapeutic anticoagulation and better outcome. In fact, fewer adverse events by better INR management may translate into improved long-term survival, as has been observed by our group [11] and others [24]. INR self-management is an evolving strategy for anticoagulation in a number of applications. It compares favorably with specialist anticoagulation clinics with regards to the INR target ranges [25, 26], it improves control of oral anticoagulation [27], and it even increases quality-of-life measures [28]. Nonetheless, patient INR self-testing is common in Europe but not in the United States. Wittkowsky and associates [29] show that costs of instruments and cartridges, and the fear that self-testing might lead to unintended self-management, are among the primary barriers [29]. These authors indicated that a reimbursement of costs would increase the frequency of INR self-testing. Indeed, the current study would not have been possible without health insurance paying for the hardware. Prevention of anticoagulation related adverse events, however, was likely to compensate for the purchase of instruments and to save health care money, whereas unintended INR self-management was not an issue. In addition, the costs associated with conventional INR management must be considered. Patients need to commute to their doctor s office, sometimes on leave from work, and medical personnel spend resources on relatively simple tasks such as anticoagulation maintenance. Therefore, patient management in remote areas of the United States and elsewhere may rather become a primary domain of INR self-testing, if reimbursement is provided by health care insurance. To be a candidate for INR self-management, however, several criteria were established by the British Society for Haematology, including involvement of clinical personnel in a formal education program of INR measurement and interpretation, documentation of INR values within the therapeutic range, and point-of-care tests of INR monitors, among others [30]. Our group closely adhered to these guidelines by offering a structured education program by dedicated medical staff [14] and by crosschecking of the devices every 6 months during the study period. In turn, a clinically relevant question arises from the fact that prosthetic heart valve recipients require lifetime anticoagulation. If so, how long can a patient pursue INR self-management? Although Siebenhofer and colleagues [31] show that INR self-management is possible in elderly patients, longer follow-up and additional studies are required. It may well be that the patients transfer their INR management to their local doctor with increasing age. Yet, maybe some patients continue to use the device, and by doing so, they may remain more self-confident and independent. Although we are currently following up several octogenarians with INR self-management, their level of acceptance has yet to be assessed. Our future studies will formally address this important topic in more detail. Clinical inferences can be derived. For instance, the choice of valve prosthesis in patients of advanced age remains a continuing challenge. There is a common yet arbitrary cutoff of 65 years of age, beyond which bioprotheses are chosen rather than mechanical valves. Although current data do not support lowering this cutoff [6], increasing it is has been controversially discussed. However, evidence accumulates that degeneration of bioprothetic valves, even in elderly and older patients with long life expectancy, would be an important factor, whereas the risk of stroke may primarily be related to patient factors [32]. Our study cannot directly support the preference of mechanical valves instead of bioprostheses; however, we demonstrated that anticoagulation-related adverse events can be minimized by INR self-management. Potentially, INR values may even be lowered because therapeutic ranges are closely met by self-testing. In fact, initial studies of our group indicate that lower INR ranges can adequately prevent bleeding events without increasing thromboembolism [23]. This study has some limitations. It was intended to result in 2 groups of equal size and of equal demographics; however, we allowed subsequent crossover between the 2 groups. The groups were thus of different age and sex, although neither variable was identified as a risk factor by our multivariate model. In addition, we previously demonstrated that the risk of adverse events did not increase although the heart valve recipients pursuing INR selfmanagement were older [23]. Furthermore, there were several instances in which patients did not report their INR

6 954 EITZ ET AL Ann Thorac Surg INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES 2008;85: levels to us during follow-up, and they had to be excluded from the current study. As expected, those patients were predominantly in the group of conventional INR management. The patients pursuing INR self-management, in contrast, had an inherent interest of reporting their INR levels and potentially seeking our advice. Thus, fewer dropouts occurred in the group of INR self-management, thereby explaining differences in sample sizes. Collectively, anticoagulation self-management can improve INR profiles up to 2 years after prosthetic valve replacement and reduces adverse events. Current indications of prosthetic rather than biologic valve implantations may be extended, if the benefit of INR self-management is shown by future studies with longer follow-up. References 1. Horstkotte D, Piper C. Improvement of oral anticoagulation therapy by INR self-management. J Heart Valve Dis 2004;13: Khan SS, Trento A, DeRobertis M, et al. Twenty-year comparison of tissue and mechanical valve replacement. J Thorac Cardiovasc Surg 2001;122: Dellgren G, David TE, Raanani E, Armstrong S, Ivanov J, Rakowski H. Late hemodynamic and clinical outcomes of aortic valve replacement with the Carpentier-Edwards Perimount pericardial bioprosthesis. J Thorac Cardiovasc Surg 2002;124: Ruel M, Kulik A, Rubens FD, et al. Late incidence and determinants of reoperation in patients with prosthetic heart valves. Eur J Cardiothorac Surg 2004;25: David TE, Ivanov J, Armstrong S, Feindel CM, Cohen G. Late results of heart valve replacement with the Hancock II bioprosthesis. J Thorac Cardiovasc Surg 2001;121: Kulik A, Bedard P, Lam BK, et al. Mechanical versus bioprosthetic valve replacement in middle-aged patients. Eur J Cardiothorac Surg 2006;30: Eitz T, Fritzsche D, Kleikamp G, Zittermann A, Horstkotte D, Korfer R. Reoperation of the aortic valve in octogenarians. Ann Thorac Surg 2006;82: Ghosh P, Holthouse D, Carroll I, Larbalestier R, Edwards M. Cardiac reoperations in octogenerians. Eur J Cardiothorac Surg 1999;15: Hammermeister K, Sethi GK, Henderson WG, Grover FL, Oprian C, Rahimtoola SH. Outcomes 15 years after valve replacement with a mechanical versus a bioprosthetic valve: final report of the Veterans Affairs randomized trial. J Am Coll Cardiol 2000;36: Koertke H, Zittermann A, Minami K, et al. Low-dose international normalized ratio self-management: a promising tool to achieve low complication rates after mechanical heart valve replacement. Ann Thorac Surg 2005;79: ; discussion Koertke H, Zittermann A, Wagner O, Koerfer R. Selfmanagement of oral anticoagulation therapy improves longterm survival in patients with mechanical heart valve replacement. Ann Thorac Surg 2007;83: Jackson SL, Bereznicki LR, Peterson GM, et al. Accuracy, reproducibility and clinical utility of the CoaguChek S portable international normalized ratio monitor in an outpatient anticoagulation clinic. Clin Lab Haematol 2004;26: Taborski U, Braun SL, Voller H. Analytical performance of the new coagulation monitoring system INRatio for the determination of INR compared with the coagulation monitor Coaguchek S and an established laboratory method. J Thromb Thrombolysis 2004;18: Koertke H, Zittermann A, Mommertz S, El-Arousy M, Litmathe J, Koerfer R. The Bad Oeynhausen concept of INR self-management. J Thromb Thrombolysis 2005;19: Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In CM M, ed. Section on microbiology. Symposia No. 2. New York, NY: Columbia University Press, 1949: Blackstone EH, Naftel TC, Turner MEJ. The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81: Hazard Package, Available at: org/heartcenter/hazard. Accessed Dec 27, Bando K, Kobayashi J, Hirata M, et al. Early and late stroke after mitral valve replacement with a mechanical prosthesis: risk factor analysis of a 24-year experience. J Thorac Cardiovasc Surg 2003;126: Ruel M, Masters RG, Rubens FD, et al. Late incidence and determinants of stroke after aortic and mitral valve replacement. Ann Thorac Surg 2004;78:77 83; discussion Hamamoto M, Bando K, Kobayashi J, et al. Durability and outcome of aortic valve replacement with mitral valve repair versus double valve replacement. Ann Thorac Surg 2003;75: 28 33; discussion Buchanan-Leel B, Levetan BN, Lombard CJ, Commerford PJ. Fixed-dose versus adjusted-dose warfarin in patients with prosthetic heart valves in a peri-urban impoverished population. J Heart Valve Dis 2002;11:583 92; discussion Adjusted-dose warfarin versus low-intensity, fixed-dose warfarin plus aspirin for high-risk patients with atrial fibrillation: Stroke Prevention in Atrial Fibrillation III randomised clinical trial. Lancet 1996;348: Koertke H, Minami K, Boethig D, et al. INR self-management permits lower anticoagulation levels after mechanical heart valve replacement. Circulation 2003;108(suppl 1):II Butchart EG, Payne N, Li HH, et al. Better anticoagulation control improves survival after valve replacement. J Thorac Cardiovasc Surg 2002;123: Cromheecke ME, Levi M, Colly LP, et al. Oral anticoagulation self-management and management by a specialist anticoagulation clinic: a randomised cross-over comparison. Lancet 2000;356: Menendez-Jandula B, Souto JC, Oliver A, et al. Comparing self-management of oral anticoagulant therapy with clinic management: a randomized trial. Ann Intern Med 2005;142: O Shea SI, Arcasoy MO, Samsa G, et al. Direct-to-patient expert system and home INR monitoring improves control of oral anticoagulation. J Thromb Thrombolysis E-pub available at q82x0g3105gg30v3/. 28. Sawicki PT. A structured teaching and self-management program for patients receiving oral anticoagulation: a randomized controlled trial. Working Group for the Study of Patient Self-Management of Oral Anticoagulation. JAMA 1999;281: Wittkowsky AK, Sekreta CM, Nutescu EA, et al. Barriers to patient self-testing of prothrombin time: national survey of anticoagulation practitioners. Pharmacotherapy 2005;25: Fitzmaurice DA, Gardiner C, Kitchen S, et al. An evidencebased review and guidelines for patient self-testing and management of oral anticoagulation. Br J Haematol 2005;131: Siebenhofer A, Rakovac I, Kleespies C, et al. Self-management of oral anticoagulation in the elderly: rationale, design, baselines and oral anticoagulation control after one year of followup. A randomized controlled trial. Thromb Haemost 2007;97: Lund O, Bland M. Risk-corrected impact of mechanical versus bioprosthetic valves on long-term mortality after aortic valve replacement. J Thorac Cardiovasc Surg 2006; 132:20 6.

7 Ann Thorac Surg EITZ ET AL 2008;85: INR SELF-MANAGEMENT AFTER PROSTHETIC VALVES DISCUSSION DR EDWARD Y. SAKO (San Antonio, TX): What was the blinding, or in both groups they recognized that they were in this type of study and what the purpose of the study was and that they were being followed for this specific thing? 955 study and who weren t followed up any longer, but when we spread them out and we analyzed them, they were equally distributed in both groups and there were essentially no adverse events in the patients that were excluded afterwards. DR SCHENK: The study inclusion criteria included the ability to use the device, so those patients may not be blinded or actually completely randomized, which is a limitation. Maybe this also explains why the patients in this self-management group were slightly younger. DR MICHAEL MACK (Dallas, TX): What percent of patients that received mechanical valves during that study period time were screened to actually enter into the study? Second, are there any reimbursement issues that would impact compliance or noncompliance in your health care system? And third, did compliance with the self-inr testing correlate with lack of adverse events? DR SCHENK: Well, I will start off with the second question. There was no reimbursement at all for any of those patients. With the first and the last question, they both relate essentially to the fact of nonblindedness to the patients. There were initially about 1000 patients screened, and I only report here 765, which is less than 80%. There were several patients who gave up the DR MACK: So is there anybody in your mind that is not a candidate for self-inr? DR SCHENK: Yes, there is. There is a requirement that patients adhere to their instructions as to the anticoagulation management. Your patients have to be able to use a cell phone or to get their data to our database. Otherwise you wouldn t consider including those. And of course, if there is any patient who is afraid of taking their life in their own hands instead of having the general practitioners do this, we would not include those also. DR MACK: But do you think you can know that ahead of time? Do you think you can predict who is going to have those issues and who isn t? DR SCHENK: After 10 or 15 years of experience with this, you can certainly predict who is going to be a good candidate or not; however, there are always patients who will drop out afterwards. That is a problem.